10.1.1 Ethical Approval
Ethical approval took approximately one year to obtain which delayed the start of data collection and reduced the duration for which the study could run. Due to the restricted time period over which the research can be carried out due to it being tied to a 3 year DDSc programme, the delay in obtaining ethical approval meant that the data collection could only run for 1 year. There were many barriers to the smooth progress of the process of obtaining ethical approval for the study.
Despite the only intervention being the measurement of patients’ heights, a full ethical review was required including an interview with the Research Ethics Committee (REC). An expert opinion and risk report was required from both a consultant radiologist and a nuclear
86 | P a g e
physicist because radiographs were being assessed as part of the research. This was despite the fact that no radiographs were being taken for the purpose of the research. We were using radiographs that had already been taken as part of the patients’ treatment planning process and were only taken when clinically indicated and justifiable, but never for the purpose of this study alone. These issues, in addition to administrative delays and IT issues with the IRAS website, contributed to the lengthy process of obtaining ethical approval.
This was further compounded by the Research Ethics Committee’s request to change the consent forms and the subsequent amendment submitted to extend recruitment and allow the following cohort of StRs to recruit patients for the study.
10.1.2 CVM not present on lateral cephalogram
In total there were 6 patients who had to be excluded from the final analysis, despite having several heights recorded, due to the vertebrae (C2, C3 and C4) not being present on the radiograph. This meant that it was not possible to use the radiographs and apply the CVM index. This highlights the importance of liaising with radiology departments to ensure that staff are aware of the correct patient positioning required when taking a lateral cephalogram to ensure that the second, third and fourth vertebrae are included on the lateral cephalogram. This issue was highlighted in the audit that was carried out and reported in Chapter 4.
10.1.3 Patients not meeting inclusion criteria
Patients who were not included in the data analysis due to not meeting the inclusion criteria totalled 4; all of these were due to the patient being too old, the ages of these 4 patients ranged from 19 to 30 years old. Despite the StRs involved being briefed on the inclusion criteria, it was inevitable that a few patients were recruited by mistake as a result of over enthusiastic recruitment. With the benefit of hindsight, it may have been worth including a checklist of the inclusion and exclusion criteria, in each surgery to serve as a reminder to the recruiting registrars; this may have reduced the number of patients recruited in error.
10.1.4 First height not recorded within 6 months of lateral cephalogram
The largest cause of patients having to be excluded from the analysis was due to lack of recorded information. The StRs involved were asked to record the height of the patient at the time their initial records were taken, and then after ethical approval had been obtained, at every appointment the patient attended. There were two main issues that were major limitations in the study. The first was that StRs did not always record the height of the patient
87 | P a g e
when they had their first cephalogram taken. This was not a problem if they recorded the height a few weeks later or even a few months later. However, there was a large number (n=41) where the height was not recorded for more than 6 months after the initial cephalogram. This meant that it was not possible to say for certain that the patient was at the CVM stage recorded on the radiograph when the first height was recorded. To ensure that this did not bias the results, any patients who had not had their first height recorded within 6 months of having their cephalogram taken were excluded. This unfortunately resulted in 41 patients being excluded from the final analysis and meant that the ideal sample size of 17 in the group, was not obtained for patients who were at CVM stages 1 and 2.
10.1.5 Patients with only one height recorded
Another large limitation of this study related to data collection, was that many patients only had one height recorded. This meant that it was not possible to calculate an annualised growth velocity for these patients and thus they had to be excluded (n=26).
Thus, in total, 77 patients had to be excluded from the final analysis, due to height measurements not being taken at the appropriate time(s). This was a substantial loss and their inclusion would have significantly increased the power of the study. It is however, a frequently encountered limitation of this type of research. Data collection depends on people without a vested interest in the study, collecting the data, and despite frequent reminders, the reality of busy clinics and being in a learning environment meant that it was inevitable that sometimes height recording was forgotten. If doing a similar study in the future, it may be of value to include a checklist to serve as a reminder to the recruiters of the steps required in the surgery in an attempt to reduce the extent of incomplete data collection.
10.1.6 Sample size to assess mandibular growth
Another limitation to this study was that only 36 of the patients who were in the final analysis, had two lateral cephalograms taken by the end of the data collection period. This meant that when measuring mandibular growth, the sample size was only 36. When this is broken down further into the different CVM stages, there were only 5 patients in the CVM stage 1 and 2 group, and only 12 patients in CVM stage 5 and 6. Due to the small numbers involved, it was not possible to show any significant differences, even though some patterns could be seen in the differences in mandibular growth at each CVM stage. The data could therefore, only be described using the means, and these should be interpreted with caution due to the small numbers involved.
88 | P a g e
10.1.7 Measuring mandibular growth
There was much discussion on how best to measure mandibular growth. Various methods which have been described in the literature, were considered as discussed in section 2.6. However, regardless of how accurate or reliable the landmarks were that were measured (Co-Gn-Go), it is debatable as to how much this triangle actually represents real mandibular growth. Although cephalometric radiographs have been used extensively in orthodontic research to assess mandibular growth, it must be remembered that a cephalogram is a magnified two-dimensional image of a three-dimensional object. There are many errors associated with landmark identification, constructing linear measurements, and the measurement of linear and angular structures.134 Baumrind and Frantz134 suggested that the observed difference as a result of therapy should be at least twice the standard deviation of the estimating error, in order to be sure that the observed difference was a biological one rather than a measurement error. This suggestion appears reasonable when it is remembered that for each linear measurement, two estimations are being made thus there are two opportunities for errors to arise. Taking this into account, together with the knowledge that condylion is particularly prone to landmark identification error, the limitations of cephalometrics must be considered very critically when assessing cephlometric data. It has been suggested that three-dimensional radiographic imaging with cone-beam computed tomography (CBCT), that allows quantitative and qualitative analysis of bone is the only thing that can precisely measure the amount and direction of mandibular growth that has occurred.83 Clearly this was not something that could be considered in this study due to ethical contraindications of exposing patients to excess radiation.
When considering the results from analysis of mandibular growth, it is important to interpret non-significant results with care. Lack of evidence of a difference in effect does not necessarily mean that there was no difference in effect.135 It is important to recognize the patterns shown whilst recognising that these means are based on very small numbers. Further efforts will be made to gather these data as the patients within the study finish their orthodontic treatment.
10.1.8 Sample
The sample was taken from the patients being taken on for treatment from the orthodontic treatment waiting list. This included only patients who had been assessed and deemed appropriate for treatment under the NHS and within a hospital department. This may reduce the generalisability of the results, but as all patients were enrolled regardless of sex, skeletal classification or dental anomalies, this should have reduced bias. All ethnicities were included in the sample and this is a key difference between this sample and the sample from
89 | P a g e
the University of Michigan Growth Study. This may be a reason as to why the growth rates at CVM stages 1,2 and 5,6 were lower than the growth rates observed in the Franchi study, if there is a substantial variation in growth velocities in different ethnic groups. However, although data regarding participants’ ethnicity was not collected, anecdotally the majority of the sample was Caucasian, so it is unlikely that the ethnic diversity of the sample was large enough to affect growth rates significantly.